The odds of having dementia at age 85 were close to 1 in 3 in the 80s; now they are 1 in 10.
— Nicolas Badre (@BadreNicolas) May 11, 2026
I don’t think we have a great explanation: better cardiovascular health, diet, and education are often mentioned. Good news nonetheless.
Carnall Farrar. (2025, March 27). Dementia… pic.twitter.com/eMizN5NcFm
But research and information I have seen elsewhere reference the 1 in 3 figure for having dementia at 85. Anticipating that this may be a cohort issue (incident rates can remain the same but the absolute number increase if the cohort itself is larger) or possibly a definitional issue, I ask Grok:
Is the decline real though? Most the claims I see are more along the lines of a one in three chance.
GroK
Yes, the decline in age-specific dementia prevalence (and incidence) is real and well-documented in high-income countries like the US and Europe.hsph.harvard.eduThe tweet from Nicolas Badre accurately reflects data from analyses like Eric Stallard's 2025 work (published in JAMA), which draws on major US surveys: the National Long Term Care Survey (1984–2004), Health and Retirement Study, and National Health and Aging Trends Study. These show age-adjusted dementia prevalence dropping sharply—by about two-thirds (around 67%) over 40 years.alzforum.orgFor example:
- In 1984, roughly 30% of people aged 85–89 had dementia.
- By 2024, that fell to about 10% (aligning with the "1 in 3 to 1 in 10" framing). alzforum.org
This is a birth cohort effect: Successive generations (e.g., those born in 1945 vs. 1895) show lower dementia rates at the same age, likely due to better education, cardiovascular health (less smoking, better blood pressure control), nutrition, and reduced vascular risks. The rate of decline has held steady at ~2.5–3% per year with no clear slowdown. dupri.duke.eduBroader EvidenceMultiple independent studies confirm this:
- Incidence (new cases) has declined ~13% per decade in Europe and the US (1988–2015), per a large meta-analysis. neurology.org
- Framingham Heart Study: ~20% drop per decade since the 1970s. nia.nih.gov
- RAND and other US data: Prevalence among those 65+ fell from ~12% in 2000 to ~8.5% in 2016. rand.org
- Similar trends in the UK/Europe, though data are sparser elsewhere. commonwealthfund.org
The "1 in 3" figure you mention is a common (and still valid) lifetime or cross-sectional risk statistic for very old age in earlier eras or without adjusting for time trends. For instance:
- Historically, ~1 in 3 people 85+ or lifetime risk estimates for older adults dying with dementia. alz.org
- It often comes from projections assuming constant age-specific rates, which don't hold up given the declines.
These older stats don't contradict the trend—they reflect what things looked like before or without accounting for improvements. Total numbers of dementia cases are still rising due to population aging (more people living to 85+), but far less dramatically than "doubling" projections suggested. Stallard's adjusted models predict only a ~25% increase by 2050 in the US, not a tsunami. alzforum.orgCaveats
- The trend is strongest in high-income Western countries; evidence is mixed or absent in many low/middle-income regions where risks may differ.
- Future trajectory isn't guaranteed—rising obesity, diabetes, and other factors could slow or reverse gains.
- Diagnosis changes, better awareness, and survival effects play some role, but the data hold after adjustments.
- Disparities persist (e.g., by race/ethnicity/education), and studies are stronger for White populations.
Overall, this is genuinely positive news: public health improvements (education, heart health, etc.) are pushing dementia risk down at any given age. It underscores prevention's power without denying the ongoing burden. For the latest, Stallard's JAMA viewpoint and related cohort studies are key references.
Man - I love this new epistemic age where complex and nuanced information is so readily available.
.jpeg)
Posts
